Federal Reserve Economic Data

The FRED® Blog

Last hired, first fired? Employment losses across age groups

Trends in 30 years of BLS employment data

The COVID-19-related recession has been especially brutal, and its effects have been unevenly distributed. One example of disparity is employment among different age groups.

The FRED graph above shows that the youngest workers were by far the worst hit: From February to April 2020, 35% of the 16- to 19-year-olds and 30% of the 20- to 24-year-olds lost their jobs. The other age categories lost a lot, too—from 11% to 16%—but much less than the youngest cohorts. While all age groups recuperated to some extent by August, the gap is still considerable.

Are the young taking one for the team just for this recession? Or have they always been first to be let go? Let’s look at the past three recessions. The second graph, which covers the period of the Financial Crisis, also shows that the youngest group was hit distinctly and persistently, while the 55 and older group actually gained employment.

The 2001 recession, which was much milder than the two we just looked at, tells the same story in the graph above: big employment losses for the youngest group and slight increases for the 45- to 54-year-old group. And below, the 1990-91 recession looks much like the 2001 recession.

All in all, it appears “normal” that the 16- to 19-year-old age group is hit hardest by recessions and that the oldest workers are largely unaffected, at least in terms of employment. The current recession is a little different in that the older groups have also been affected, just not as much as the younger groups.

How these graphs were created: Start from Table A-9 of the Current Population Survey, select the series you want shown, and click “Add to Graph.” From the “Edit Graph” panel, select units “Index” with the start of the current recession and click “Apply to All.” Adjust the sample period. For the other graphs, adjust the index date and sample periods.

Suggested by Christian Zimmermann.

View on FRED, series used in this post: LNS12000012, LNS12000036, LNS12000089, LNS12000091, LNS12000093, LNS12024230

Most unemployment measures are declining…

...while long-term unemployment is still rising

Many of us follow the unemployment rate closely, even more so since the pandemic began. But there are many definitions of unemployment, which depend on how people are attached to the labor force. To learn more, see this earlier blog post and this conversational account of unemployment measures.

Today’s FRED graph shows the recent evolution of 6 measures of unemployment. All increased dramatically, but not uniformly: The lines didn’t move in a parallel fashion—that is, the distance between them didn’t remain constant. Rather, the lines fanned out, showing that it wasn’t one particular type of unemployment that was responsible for the overall surge.

One detail worth noting, though, is that the long-term unemployed, which by definition take some time to accumulate, are still increasing, while all other unemployment groups are decreasing.

As of August, the long-term unemployed made up 5.1% of the labor force. And if the long-term unemployment rate stays high, the general unemployment rate must stay high, too. If the previous recession is any indication, reducing long-term unemployment may take a long time. Adjust the graph sliders to include the time period of the previous recession, and you’ll see what we mean.

How this graph was created: From the Alternative Measures of Labor Underutilization release table (A-15) from the Bureau of Labor Statistics’ Employment Situation release, select all (seasonally adjusted) series and click “Add to Graph.” Adjust the sample period as you wish.

Suggested by Christian Zimmermann.

View on FRED, series used in this post: U1RATE, U2RATE, U4RATE, U5RATE, U6RATE, UNRATE

How much commuting time are we saving by working from home?

A back-of-the-envelope calculation of pandemic-related changes

The FRED Blog has looked at the wide range of commuting times across U.S. cities and counties, as well as the impact of shorter commutes on employment and happiness.

Given that many employees have been working from home during the COVID-19 pandemic, we’ll try to gauge the potential number of hours per week that are no longer spent commuting to work. Clearly, not every employee is working from home these days. So this is a “back-of-the-envelope” calculation, which uses available information to approximate answers to very complex questions.*

First, we use the latest data (which is from 2018) on the average daily commuting time for three suburban counties:

  • 29.57 minutes in DuPage County, IL
  • 24.30 minutes in St. Louis County, MO
  • 32.18 minutes in Fairfax, VA

Next, we multiply those average commuting times by the number of employed persons in those counties. Then we divide each figure by 60 to transform minutes to hours and multiply that by 5 to account for a 5-day workweek.

The FRED graph above shows the potential number of hours per week spent on commuting that are being saved by working from home: Between April and July, that potential weekly time savings in each county ranges from 1 million to 1.5 million hours.

*The physicist Enrico Fermi used this method to great effect in his research and teaching to get rough orders of magnitude. We wonder what Professor Fermi would have been able to accomplish if he had access to FRED… He received a Nobel Prize in physics in 1938 and passed in 1954, long before FRED came to be.

How this graph was created: Search for and select “Employed Persons in DuPage County, IL.” From the “Edit Graph” panel, use the “Add Line” tab to search for and select “Employed Persons in St. Louis County, MO” and “Employed Persons in Fairfax County, VA.” Next, click on “1Y” above the graph to display the last 12 observations. Next, customize each line by applying the formula (a*average commuting time/60)*5. For Line 1, that formula is (a*29.57/60)*5. Last, from the “Edit Graph” panel, click on the “Format” tab and select line colors and mark types to taste.

Suggested by Diego Mendez-Carbajo.

View on FRED, series used in this post: LAUCN170430000000005, LAUCN291890000000005, LAUCN510590000000005


Back to Top